Crop&Scale系¶
- RandomCrop
- ShiftScaleRotate (Rotate系に記載)
- CenterCrop
- PadIfNeeded
- Crop
- RandomScale
- LongestMaxSize
- SmallestMaxSize
- Resize
- RandomSizedCrop
- RandomCropNearBBox
- RandomSizedBBoxSafeCrop
In [1]:
from albumentations import (
RandomCrop,
CenterCrop,
PadIfNeeded,
Crop,
RandomScale,
LongestMaxSize,
SmallestMaxSize,
Resize,
RandomSizedCrop,
RandomCropNearBBox,
RandomSizedBBoxSafeCrop
)
from utils import *
In [2]:
data = get_data()
image = data['image']
image_seg = data['seg_img']
mask = data['mask']
image_det = data['det_img']
bboxes = data['bboxes']
RandomCrop¶
In [3]:
"""Crop a random part of the input.
Args:
height (int): height of the crop.
width (int): width of the crop.
p (float): probability of applying the transform. Default: 1.
Targets:
image, mask, bboxes, keypoints
Image types:
uint8, float32
"""
trans = RandomCrop(height=128, width=128, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
CenterCrop¶
In [4]:
"""Crop the central part of the input.
Args:
height (int): height of the crop.
width (int): width of the crop.
p (float): probability of applying the transform. Default: 1.
Targets:
image, mask, bboxes, keypoints
Image types:
uint8, float32
Note:
It is recommended to use uint8 images as input.
Otherwise the operation will require internal conversion
float32 -> uint8 -> float32 that causes worse performance.
"""
trans = CenterCrop(height=128, width=128, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
PadIfNeeded¶
In [5]:
"""Pad side of the image / max if side is less than desired number.
Args(非公式):
min_height (int): minimam height of padding image. If height of original image is greater than min_height,
height of output is not changed. Default: 1024.
min_width (int): minimam width of padding image. If width of original image is greater than min_width,
width of output is not changed. Default: 1024.
border_mode (OpenCV flag): flag that is used to specify the pixel extrapolation method. Should be one of:
cv2.BORDER_CONSTANT, cv2.BORDER_REPLICATE, cv2.BORDER_REFLECT, cv2.BORDER_WRAP, cv2.BORDER_REFLECT_101.
Default: cv2.BORDER_REFLECT_101
p (float): probability of applying the transform. Default: 1.0.
value (list of ints [r, g, b]): padding value if border_mode is cv2.BORDER_CONSTANT.
Targets:
image, mask, bbox, keypoints
Image types:
uint8, float32
Comment:
min_heightやmin_widthよりも入力画像が大きければ、高さや幅は変わりません。
もしも、それより大きければ、border_modeに従うPaddingがなされて拡大します。
"""
trans = PadIfNeeded(min_height=1024, min_width=1024, border_mode=2, value=None, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
Crop¶
In [6]:
"""Crop region from image.
Args:
x_min (int): minimum upper left x coordinate
y_min (int): minimum upper left y coordinate
x_max (int): maximum lower right x coordinate
y_max (int): maximum lower right y coordinate
Targets:
image, mask, bboxes
Image types:
uint8, float32
Comment:
任意のサイズにCropします。引数は与える画像のサイズ内にしてください。
"""
trans = Crop(x_min=0, y_min=0, x_max=128, y_max=128, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
RandomScale¶
In [7]:
"""Randomly resize the input. Output image size is different from the input image size.
Args:
scale_limit ((float, float) or float): scaling factor range. If scale_limit is a single float value, the
range will be (1 - scale_limit, 1 + scale_limit). Default: 0.1.
interpolation (OpenCV flag): flag that is used to specify the interpolation algorithm. Should be one of:
cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4.
Default: cv2.INTER_LINEAR.
p (float): probability of applying the transform. Default: 0.5.
Targets:
image, mask, bboxes, keypoints
Image types:
uint8, float32
Comment:
scale_limitに1つの値を与えた場合は、(1 - scale_limit, 1 + scale_limit)の範囲、タプルを与えた場合は、
その範囲内からランダムに値が選ばれます。
注意点としては、1+選ばれた数値の割合だけ拡大縮小されるので、1.0が選ばれれば、画像は2.0倍、-0.5が選ばれれば0.5倍、
つまり半分のサイズに画像が変換されます。
"""
trans = RandomScale(scale_limit=(-0.5, -0.5), interpolation=1, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
LongestMaxSize¶
In [8]:
"""Rescale an image so that maximum side is equal to max_size, keeping the aspect ratio of the initial image.
Args:
p (float): probability of applying the transform. Default: 1.
max_size (int): maximum size of the image after the transformation
Targets:
image, mask, bboxes
Image types:
uint8, float32
Comment:
入力画像の縦横比を維持しながら、最大の辺の長さをmax_sizeに与えた画像サイズに変換します。
"""
trans = LongestMaxSize(max_size=512, interpolation=1, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
SmallestMaxSize¶
In [9]:
"""Rescale an image so that minimum side is equal to max_size, keeping the aspect ratio of the initial image.
Args:
p (float): probability of applying the transform. Default: 1.
max_size (int): maximum size of smallest side of the image after the transformation
Targets:
image, mask, bboxes
Image types:
uint8, float32
Comment:
入力画像の縦横比を維持しながら、最小の辺の長さをmin_sizeに与えた画像サイズに変換します。
"""
trans = SmallestMaxSize(max_size=512, interpolation=1, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
Resize¶
In [10]:
"""Resize the input to the given height and width.
Args:
p (float): probability of applying the transform. Default: 1.
height (int): desired height of the output.
width (int): desired width of the output.
interpolation (OpenCV flag): flag that is used to specify the interpolation algorithm. Should be one of:
cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4.
Default: cv2.INTER_LINEAR.
Targets:
image, mask, bboxes
Image types:
uint8, float32
Comment:
与えた任意の height(縦) x width(横)サイズに画像を変換します。
"""
trans = Resize(height=128, width=256, interpolation=1, always_apply=False, p=1)
imshow2(image, transforms=trans)
imshow4(image_seg, mask, transforms=trans)
RandomSizeCrop¶
In [11]:
"""Crop a random part of the input and rescale it to some size.
Args:
min_max_height ((int, int)): crop size limits.
height (int): height after crop and resize.
width (int): width after crop and resize.
w2h_ratio (float): aspect ratio of crop.
interpolation (OpenCV flag): flag that is used to specify the interpolation algorithm. Should be one of:
cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4.
Default: cv2.INTER_LINEAR.
p (float): probability of applying the transform. Default: 1.
Targets:
image, mask, bboxes, keypoints
Image types:
uint8, float32
Comment:
min_max_heightの範囲内でランダムにCropの高さが選ばれ、その高さ x 与えたw2h_ratioがCropの幅として設定されます。
例えば、Cropの高さが128とし、w2h_ratioに2.0を与えると、Cropサイズは 128x256 になります。
Cropされた部分は、引数として与えた height x width のサイズに変換されます。
"""
trans = RandomSizedCrop(min_max_height=(128, 244), height=244, width=244, w2h_ratio=1.0,
interpolation=1, always_apply=False, p=1.0)
imshow2(image, transforms=trans)
trans = RandomSizedCrop(min_max_height=(400, 700), height=720, width=1280, w2h_ratio=1280/720,
interpolation=1, always_apply=False, p=1.0)
imshow4(image_seg, mask, transforms=trans)
RandomCropNearBBox¶
In [12]:
"""Crop bbox from image with random shift by x,y coordinates
Args:
max_part_shift (float): float value in (0.0, 1.0) range. Default 0.3
p (float): probability of applying the transform. Default: 1.
Targets:
image
Image types:
uint8, float32
Comment:
与えたbbox近くをランダムにシフトした部分をCropした画像に変換します。
注意点としては、bboxは複数与えられません。1つだけ与えてください。加えて、このbboxだけは、alubmentationsのフォーマット([0,1])ではなく、
Pascal VOCのフォーマット[x_min, y_min, x_max, y_max]で渡してください。
callableには、image, cropping_bboxを与えてください。
"""
img=putBbox(image_det, [bboxes[5]])
trans = RandomCropNearBBox(max_part_shift=0.3, always_apply=False, p=1.0)
tsf_dict = trans(image=img, cropping_bbox=bboxes[5])
fig, ax = plt.subplots(nrows=2, ncols=1, figsize=(10, 10))
ax[0].tick_params(labelbottom="off",bottom="off")
ax[0].tick_params(labelleft="off",left="off")
ax[1].tick_params(labelbottom="off",bottom="off")
ax[1].tick_params(labelleft="off",left="off")
ax[0].set_title('Original [{} x {}]'.format(img.shape[0], img.shape[1]))
ax[1].set_title('Transformed [{} x {}]'.format(tsf_dict['image'].shape[0], tsf_dict['image'].shape[1]))
ax[0].imshow(img.astype('uint8'))
ax[1].imshow(tsf_dict['image'].astype('uint8'))
fig.show()
RandomSizedBBoxSafeCrop¶
In [13]:
"""Crop a random part of the input and rescale it to some size without loss of bboxes.
Args:
height (int): height after crop and resize.
width (int): width after crop and resize.
erosion_rate (float): erosion rate applied on input image height before crop.
interpolation (OpenCV flag): flag that is used to specify the interpolation algorithm. Should be one of:
cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4.
Default: cv2.INTER_LINEAR.
p (float): probability of applying the transform. Default: 1.
Targets:
image, mask, bboxes
Image types:
uint8, float32
Comment:
バウンディングボックス情報とともに画像をランダムにscale&cropします。
callableには、image, bboxesを与えてください。
"""
# Pascal VOC フォーマットから albumentations フォーマットにバウンディングボックスを変換します。
bbox_albu = albu.augmentations.bbox_utils.convert_bboxes_to_albumentations(bboxes=bboxes, source_format='pascal_voc',
rows=image_det.shape[0], cols=image_det.shape[1], check_validity=False)
trans = RandomSizedBBoxSafeCrop(height=256, width=256, erosion_rate=0.2, interpolation=1, always_apply=False, p=1.0)
tsf_dict = trans(image=image_det, bboxes=bbox_albu)
tsf, tsf_bb = tsf_dict['image'], tsf_dict['bboxes']
# albumentationsのフォーマットからPascal VOCのフォーマットに変換 (注意: rows, colsには変換後の画像サイズを入力する。)
tsf_bb = albu.augmentations.bbox_utils.convert_bboxes_from_albumentations(bboxes=tsf_bb, target_format='pascal_voc',
rows=tsf.shape[0], cols=tsf.shape[1], check_validity=False)
fig, ax = plt.subplots(nrows=2, ncols=1, figsize=(10, 10))
ax[0].tick_params(labelbottom="off",bottom="off")
ax[0].tick_params(labelleft="off",left="off")
ax[1].tick_params(labelbottom="off",bottom="off")
ax[1].tick_params(labelleft="off",left="off")
ax[0].set_title('Original [{} x {}]'.format(image_det.shape[0], image_det.shape[1]))
ax[1].set_title('Transformed [{} x {}]'.format(tsf.shape[0], tsf.shape[1]))
ax[0].imshow(putBbox(image_det, bboxes,thickness=3).astype('uint8'))
ax[1].imshow(putBbox(tsf, tsf_bb, thickness=2).astype('uint8'))
fig.show()
